...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Optical engineering >Multiple incident angles and wavelengths identifiable proximity optical communication scheme utilizing the Talbot effect and deep neural network
【24h】

Multiple incident angles and wavelengths identifiable proximity optical communication scheme utilizing the Talbot effect and deep neural network

机译：利用Talbot效应和深度神经网络的多个入射角和波长可识别的近距离光通信方案

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

A lensless light field image sensor configuration is presented to realize a proximity light communication system from multi-light-emitting diode (LED) clusters with different incident angles. The light field image sensor is fabricated on a standard complementary metal-oxide-semiconductor (CMOS) process and utilizes the Talbot effect. Two metal layers are stacked above the photodiode acting as diffraction gratings at the top and a masking layer at the bottom placed at Talbot depth to either pass or block the incident light. In particular, we introduce a proximity optical communication scheme to achieve dual light communication from different incident angles with various wavelengths. To simplify the classification between different combinations of incident angles and wavelengths, a deep neural network (DNN) is utilized. We reduced the number of light field image sensors by 50% and increased the transmission data bits per clock by 114% compared with our previous work. A test environment for deciphering three incident angles and three types of wavelength combinations is devised. For each incident angle, three LEDs, which are red, green, and blue (RGB), are located at precomputed locations. A 3 × 3 chip LED array consisting of three sets of RGB LEDs is used to transmit signals, and a light field image sensor chip acts as a receiver 3.5 cm away. A range of 46 different types of blink combinations are tested from this single LED array considering up to dual lights communication. The light field image sensor chip consists of six photodiodes with different grating pairs. In particular, three types of grating layer combinations with the photodiodes can detect nine different single blinks and 36 different dual blink combinations, thus totaling 46 types of spatial signals. We achieved a maximum data rate of 251 kbit/s with the proposed prototype at a 45.5-kHz sample rate. The proposed proximity optical communication scheme had power dissipations of 30.6 mW on a 1.8-V power supply for the sensor chip and 27.5 mW on a 3.3-V power supply for the LED array. This can be applied in the low-power data input/output communication between a compact cluster of chips.
机译：提出了一种无透镜光场图像传感器配置，以通过具有不同入射角的多发光二极管（LED）集群实现近距离光通信系统。光场图像传感器采用标准的互补金属氧化物半导体（CMOS）工艺制造，并利用了Talbot效应。两层金属层堆叠在光电二极管上方，顶部用作衍射光栅，底部则用作Talbot深度的掩模层，以通过或阻挡入射光。特别是，我们引入了一种近距离光通信方案，以实现从具有不同波长的不同入射角的双光通信。为了简化入射角和波长的不同组合之间的分类，使用了深度神经网络（DNN）。与我们以前的工作相比，我们将光场图像传感器的数量减少了50％，并将每个时钟的传输数据位增加了114％。设计了一种用于解密三种入射角和三种波长组合的测试环境。对于每个入射角，红色，绿色和蓝色（RGB）的三个LED位于预先计算的位置。由三组RGB LED组成的3×3芯片LED阵列用于传输信号，光场图像传感器芯片充当3.5厘米远的接收器。从这个单个LED阵列考虑到双灯通信，测试了46种不同类型的闪烁组合。光场图像传感器芯片由六个具有不同光栅对的光电二极管组成。特别地，具有光电二极管的三种类型的光栅层组合可以检测九种不同的单眨眼和36种不同的双眨眼组合，因此总计46种类型的空间信号。所提出的原型在45.5 kHz的采样率下实现了251 kbit / s的最大数据速率。拟议的邻近光通信方案在传感器芯片的1.8 V电源上的功耗为30.6 mW，在LED阵列的3.3 V电源上的功耗为27.5 mW。这可以应用于紧凑型芯片簇之间的低功耗数据输入/输出通信。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号